Over the past year, I have been building an AI assistant for attorneys called Lawful Good. The product is designed specifically for independent attorneys.

It grew out of work I had been doing to understand how to build an effective Retrieval Augmented Generation (RAG) AI system. As I was doing this work, I thought about who would make the best use of this, and attorneys leapt out to me as the ones who could most benefit from a tool like this. Attorneys need document analysis, document generation, research, and data privacy. These are all things that the system I was building could do.

I began testing the application with our family’s trust and estate lawyer. At first, it didn’t really click with him. He would use it occasionally for research and some document analysis. After a couple of months, he decided to go all in, and devote real time and energy into seeing what it could do. This is when things really took off for him. He started to understand what the AI agent could and could not do. More importantly, he started to understand how important context and prompt crafting is to the quality of responses. He told me it took him about 6 months to finally really “get” it, and would start to use it on a daily basis. This is really interesting, because for me, there has been no ramp up time. But, this is because I have been involved in AI before ChatGPT was even released, and I have been using chatbots since before they were really very useful at all. But, most lawyers have been hesitant to use AI for their practice. Therefore, when they do try it, there is the inevitable ramp up time because everything is so new. However, the difference between this kind of ramp up and the learning curves of traditional software is quite different. With an AI assistant, there aren’t a lot of buttons, menus, dials, and configurations to learn and memorize. Instead, the way you ask questions and instruct the agent can make all the difference. In addition, understanding how AI systems work - how they simulate memory, the extent of their context windows, and when to sense that context rot has begun to affect output all are very subtle, non-deterministic skills to learn. In a way, learning how to get the most out of an AI system is just as much of an art as it is a skill.

Over the span of 2025, I gradually grew the Pilot Program for Lawful Good to several more attorneys who were interested in experimenting with AI for their practice. In my conversations with these users, I started to understand a common theme. All of these individuals were very competent attorneys. They had also chosen to go solo for a reason. They could have worked in a larger firm, but instead chose to go out on their own. This is when the brand for Lawful Good began to take shape. Instead of emphasizing product features and capabilities, I started to position the product around the deeper reasons these lawyers were shaping their own careers: to have greater autonomy, to be free of the office politics that plague Big Law, to chart their own course. In a word, freedom. At this point, I realized the product was going to be as much about lifestyle as it would be about raw capabilities. My mission became more clear. At this time, there were many companies building legal AI systems. But all of them appeared to be targeting Big Law, or at the smallest, medium-sized law firms. There were not many offerings out there targeting solo attorneys, and those that were did not seem to have simple, transparent pricing. I found a statistic that said that only around 4% of solo attorneys were using AI as a part of their practice. Clearly, they are an underserved market. I decided that this would be my niche. The Lawful Good product would be built specifically around the needs of independent attorneys, and the promise would be simplicity: simple pricing, simple onboarding, a simple user experience that immediately enhances the productivity of the user.

What’s with the name?

You might be wondering where I got the name. When I was casting about for a name for this product, I had been thinking a lot about AI alignment (an extremely important subject, but one too large to get into here). You may or may not be familiar with the alignment system in the role-playing game Dungeons & Dragons. The alignment system in the game is a simple method for categorizing different personality types using two dimensions: lawful to chaotic, and good to evil. Therefore, “Lawful Good” seemed like it had two characteristics that one would definitely want their legal AI assistant to have: "Lawful" and "Good" (Chaotic Evil seemed like a poor marketing choice). Nerdy? Yes. Distinctive? Also, yes - so why not?

Give it a try

If you are an attorney, and you are interested in trying out AI as part of your practice, you can read more about it at: https://lawfulgood.us. You can sign up and start using it right away at https://app.lawfulgood.us